Weather Variability based Trends and Yield Forecasting using Deep Learning in 3 A Agroclimatic Zone

Abstract

Economic growth and prosperity of a nation are inextricably linked to the agricultural sector. newlineWeather variability plays a crucial role in crop yield outcomes. In the compass of agriculture, newlineclimate and other environmental changes are one of the main challenges. Understanding the newlinetrends in weather variables and their impact on crop yield can aid in effective agricultural newlineplanning and decision-making. This study investigates changes in rainfall, temperature and newlineother weather variables in Jaipur district over a 37-year period and at the same time aims to newlineleverage the power of deep learning algorithms to capture complex relationships between newlineweather variables and crop productivity, enabling accurate yield predictions of major crops in newlinethe 3A agroclimatic zone of Rajasthan. The analysis reveals a non-significant increasing trend newlinein minimum temperature and a significant increase in maximum temperature over time, newlineconfirmed by the Mann-Kendall trend test. Rainfall exhibits a non-significant decreasing trend newlineduring the study period. Through comparative analysis of various Machine Learning newlinealgorithms, the study identifies Random Forest, a supervised learning model as the best- newlineperforming algorithm for yield prediction achieving an overall accuracy of 92.3%. This newlineapproach facilitates optimal forecasting, assisting farmers and policymakers in crop production newlineand management planning. Furthermore, the study provides scientific recommendations based newlineon the findings, benefiting farmers, policymakers, and other stakeholders. By addressing newlineenvironmental and agricultural challenges in this semi-arid region of Rajasthan facing climate newlinechange issues, the results contribute to sustainable solutions and agricultural advancements. newline